Talked about this week

Terms used by more blogs in the last 7 days than in the 12 weeks before, ranked by breadth times surprise. Each blog counts once per term.

News outlets and publications are left out by default: a quarter of the raw list was newspapers.

#
Term
Blogs
Baseline
Lift
Headline
1
flydubai
38
2
×68.8
In phone call, PM Modi lauds courage of flydubai pilot, hears his account of mid-air incidentoriginal ↗
thehindubusinessline.com
2
cornell
47
20
×12.2
Cornell 7 shows colleges can't probe sex assaultsoriginal ↗
archive.li
3
halloween
65
106
×3.3
Halloween 2026: Here's your guide to haunted houses in the Philadelphia regionoriginal ↗
6abc.com
4
tennessee
48
65
×4.0
Is Christa Pike still alive: Convicted killer in critical condition after botched execution in Tennessee, attorneys sayoriginal ↗
6abc.com
5
injured
43
91
×2.5
Nushrratt Bharuccha injured in Bali, suffers spine fracture; team plans air ambulance to Mumbaioriginal ↗
dnaindia.com
6
rape
35
58
×3.2
What we know about the Cornell University rape investigationoriginal ↗
bworldonline.com
7
pope
36
63
×3.1
Sub-Rural, #64 Pope.Loriginal ↗
badatsports.com
Method

How a term is ranked

score = breadth × lift
breadth = blogs using the term this window / blogs publishing this window
lift = (breadth this window / typical breadth in a window-length slice of the baseline)
       / the same ratio for the median term

A term has to be at least two and a half times as broad as in a typical earlier window of the same length, measured against how much broader the median term reads (recent weeks are held more completely than older ones), and used by a floor of blogs rather than a handful, so an ordinary word drifting from one week to the next does not read as news. Function words, generic headline verbs and the calendar are dropped before counting: months and weekdays trend every week by definition. A source counts only if it published before the window opened, so a blog new to the corpus cannot read as a spike.